RESEARCH ARTICLE

Biodiversity in Hudson River shore zones: influence of shorelinetype and physical structure

David L. Strayer Stuart E. G. Findlay Daniel Miller

Heather M. Malcom David T. Fischer Thomas Coote

Received: 24 June 2011 / Accepted: 2 February 2012

Springer Basel AG 2012

Abstract The shore zones of the Hudson River, like those

of many developed waterways, are highly varied, containing

a mix of seminatural and highly engineered shores. Our goal

was to document the biodiversity supported by different

kinds of shore zones in the Hudson. We chose six common

types of shore zones, three of them natural (sand, uncon-

solidated rock, and bedrock), and three of them engineered

(riprap, cribbing, and bulkheads). We measured selected

physical characteristics (shore zone width, exposure, sub-

strate roughness and grain size, shoreline complexity) of

three examples of each of these shore types, and also sampled

communities of terrestrial plants, fishes, and aquatic and

terrestrial invertebrates. Community composition of most

taxa differed across shore types, and frequently differed

between wide, sheltered shores and narrow, exposed shores.

Alien plant species were especially well represented along

engineered shores. Nevertheless, a great deal of variation in

biological communities was not explained by our six-class

categorization of shore zones or the physical variables that

we measured. No single shore type supported the highest

values of all kinds of biodiversity, but engineered shore

zones (especially cribbing and bulkheads) tended to have

less desirable biodiversity characteristics than natural

shore zones.

Keywords Shorelines Littoral zone Estuary Fish Macroinvertebrates Riparian zone Vegetation

Introduction

The shore zone of aquatic ecosystems (the region adjoining

the shoreline in which direct interactions tightly link the

terrestrial and aquatic ecosystems) is ecologically impor-

tant, highly modified and understudied (e.g., Airoldi and

Beck 2007; McLachlan and Brown 2006; National

Research Council 2007; Strayer and Findlay 2010). As a

consequence of the typically high heterogeneity and large

inputs of organic matter to the shore zone, both aquatic and

terrestrial biodiversity often are high in the shore zone,

including many species not found in adjacent habitats (e.g.,

Ward et al. 1999; Sabo et al. 2005; Strayer and Findlay

2010; Kennedy and Turner 2011), and rates of primary

production, respiration, and other biogeochemical pro-

cesses can be very high (e.g., Wetzel 1990; Polis and Hurd

1996; Coupland et al. 2007; Strayer and Findlay 2010).

However, human activities such as agriculture, urbaniza-

tion, commercial navigation, and recreation have been

focused on shore zones for millennia, so many shore zones

have been highly modified (e.g., Tockner and Stanford

2002; Scholten et al. 2005; Strayer and Findlay 2010). In

particular, many shore zones now contain structures such

as walls and revetments designed to protect valuable

waterfront property. Consequently, the biodiversity and

ecological function of many shore zones has been greatly

reduced (e.g., Strayer and Findlay 2010; Kennedy and

Turner 2011).

D. L. Strayer (&) S. E. G. Findlay H. M. Malcom D. T. Fischer

Cary Institute of Ecosystem Studies,

P.O. Box AB, Millbrook, NY 12545, USA

e-mail: strayerd@caryinstitute.org

D. Miller

Hudson River National Estuarine Research Reserve,

New York State Department of Environmental Conservation,

Hudson River Estuary Program, Staatsburg, NY, USA

T. Coote

Department of Environmental Conservation,

University of Massachusetts, Amherst, MA 01003, USA

Aquat Sci

DOI 10.1007/s00027-012-0252-9 Aquatic Sciences

123

Despite their importance and imperiled condition, shore

zones may be understudied because they are claimed by

neither aquatic nor terrestrial ecologists. Further, the

extremely high heterogeneity of the shore zone makes it

difficult to design sampling schemes and apparatus that are

efficient throughout the full range of conditions encoun-

tered. As a result, detailed information on the biodiversity

and ecological functioning of shore zones often is lacking,

especially along fresh waters.

Nevertheless, information about the biodiversity and

ecological function of different kinds of shore zones is

necessary for wise management of these important eco-

systems. A challenge for shore zone management is to

preserve human uses of the shore zone while retaining or

improving ecological function. To reach this goal, we will

have to understand what characteristics of shore zones

determine their biodiversity and ecological functioning.

An extensive review of the literature led to the conclu-

sion that shore zone biodiversity depends chiefly on the

physical energy regime, geologic (or anthropogenic)

structure, the hydrologic regime, nutrient inputs, and cli-

mate (Strayer and Findlay 2010). If we are thinking of

managing a particular site, physical structure often will be

the easiest of these factors to manipulate. In particular, high

biodiversity often is associated with high physical hetero-

geneity, so it may be desirable to design sites or manage for

high physical heterogeneity.

Here, we report on surveys of biodiversity in the shore

zone of the freshwater tidal Hudson River. The goals of our

study were to (1) describe the biological communities that

occupy the shore zones of the freshwater tidal Hudson River;

(2) test whether those biological communities differ among

different types of shore zones, whether natural and engi-

neered; and (3) test whether the physical characteristics of

shore zones (e.g., roughness, slope, exposure) can be used to

predict the attributes of their biological communities. Spe-

cifically, we hypothesized that high biodiversity would be

associated with high physical complexity of shore zones, and

that engineered shore zones would be physically simpler and

therefore support lower biodiversity than natural shore

zones. Instead of focusing on a single taxonomic group, we

deliberately chose to study several taxa, both aquatic and

terrestrial. We hope that this information will help to better

manage shore zones along the Hudson and elsewhere.

Methods

The study area

The study area is the freshwater tidal portion of the Hudson

River (Fig. 1), which extends from RKM 100 (i.e., river

kilometer 100 as measured from the Battery at the southern

end of Manhattan in New York City) to the head of tide at

RKM 248 in Troy. The entire study area is subject to twice-

daily tides with an average range of 0.81.6 m; tidal range

does not diminish upriver and is as large at Troy as at

Manhattan (Geyer and Chant 2006). Because daily tidal

flows are much greater than typical freshwater flows

(Geyer and Chant 2006), water levels along the Hudsons

shores are determined more by tides than by freshwater

flows (Strayer and Findlay 2010; Fig. 17), and the Hudson

does not have the large, well-defined floodplain charac-

teristic of typical alluvial rivers. The average water depth is

*8 m and the average channel width is *900 m. TheHudson is used by ocean-going ships as far as Albany

(RKM 226), and smaller commercial ships all the way to

Troy, and is heavily used by recreational boaters

throughout the study area. Wakes from this traffic probably

are important forces on shore zones throughout the study

area.

The water is moderately hard, turbid, and fertile (typical

growing-season values: calcium = 25 mg/L, NO3-N =

0.5 mg/L, PO4-P = 28 lg/L, Secchi transparency = 1 m;Caraco et al. 1997; Simpson et al. 2006). Large beds of the

Fig. 1 Location of the study sites (black circles and italic labels) inthe freshwater tidal Hudson River in eastern New York, USA. The

first letter of the site code gives the section of river (L lower,M middle, U upper), and the last two letters give the shore type (BHbulkhead, BR bedrock, CR cribbing, RO natural rock, RR riprap, SAsand)

D. L. Strayer et al.

123

macrophytes Vallisneria americana (submersed) and Trapa

natans (floating-leaved) are common (Nieder et al. 2004).

The fauna is dominated by species typical of warm fresh

waters, although diadromous fishes (e.g., Alosa spp.) and

several typically brackish-water fishes and invertebrates are

common in the freshwater estuary (e.g., Simpson et al. 1985).

We defined the shore zone as the region extending from

1 m vertically below mean low water to 1 m vertically

above mean high water. Although admittedly arbitrary, this

definition is consistent with previous definitions (reviewed

by Strayer and Findlay 2010) that delimit the shore zone as

a region of strong interactions between land and water, and

was practical to apply in the field. Note that this definition

of the shore zone includes areas that are usually under

water, areas that are usually exposed to the air, and areas

that are intermittently inundated.

Study design

We studied six types of shore zones, three of them engi-

neered (riprap, cribbing, and steel or concrete bulkheads)

and three of them without obvious signs of human engi-

neering (bedrock, natural unconsolidated rock, and sand).

Even these natural shore zones often showed some signs

of past human activity. Definition of shore types was based

on predominant conditions at and just above the high water

mark. Riprap shores are revetments constructed of large

stone (typically [50 cm in diameter), cribbing shoresconsist of wooden pilings (typically *25 cm diameter)back-filled on the land side with crushed stone 1525 cm in

diameter, and bulkheads are vertical walls or revetments

made of steel or concrete. These engineered structures

typically extend from well below mean low water to well

above mean high water, although some parts of the crib-

bing are overtopped at mean high water. Designation of

shore types and their locations were based on a complete

GIS inventory of both shores of RKM 50248 of the

Hudson in 2006 (D. Miller, unpublished).

We selected one example of each type of shore zone in

three reaches of the Hudson (Fig. 1): upper (RKM

213247), middle (RKM 151213), and lower (RKM

100151). Blocking by reach was intended to control for

northsouth differences in biota along the length of the

river. We selected a random point in each river reach, then

chose the closest example of each shore type to that ran-

dom point that met the following criteria: it was at least

130 m long (i.e., 100 m of shore to be studied plus at least

a 15 m buffer on each end) and was not interrupted by a

tributary mouth or any major human structure (e.g., a

marina). In two cases where a particular shore type was

rare in a section of the river, we had to relax these stan-

dards and choose a slightly shorter reach of shore (the

shortest reach was 75 m long).

Physical variables

We measured several physical properties of each site: shore

zone width, sediment grain size, exposure, substrate

roughness, and shoreline complexity. We measured shore

zone width as the distance along a line perpendicular to the

shoreline from 1 m vertically above the high-water mark to

1 m vertically below the low-water mark. At a few sites

with very shallow slopes, we measured the underwater

slope only to a point 60 m from the shoreline, where the

water was less than 1 m deep and extrapolated the slope

from the low water mark to the 1 m depth contour. We

repeated this procedure three times at each site: at the

upriver end, middle, and downriver end of the site. These

estimates were then combined into a single estimate of

shore zone width for each site.

We used the modified pebble count method of Lazor-

chak et al. (1998) to estimate sediment grain size in the

intertidal zone. Briefly, we laid out a 100-m long tape along

the low-water line, the mean-water line, and the high-water

line, and categorized the sediment at 1 m intervals along

the line as 0 = silt or clay, 1 = sand, 2 = sand to the size

of a marble (16 mm), 3 = size of a marble to a tennis ball

(64 mm), 4 = size of a tennis ball to a basketball

(250 mm), 5 = size of a basketball to 1 m diameter,

6 = 14 m diameter, and 7 = larger than 4 m diameter.

We averaged these 300 measurements to get a single

estimate of average sediment grain size for each site.

Sediment grain size of subtidal sediments was estimated

from handheld cores (5 cm diameter) taken at depths of

0.33 and 0.67 m below the low-water mark. Three cores

were taken at each depth at each site. We dried each

sample, weighed it, than wet-sieved the sample through

brass soil sieves of 0.25, 1, and 4 mm mesh, and measured

the dry mass of sediment retained on each sieve. We then

calculated an index of grain size as [mass retained on

0.25-mm mesh sieve ? 2 9 (mass retained on 1-mm mesh

sieve) ? 3 9 (mass retained on 4-mm mesh sieve)]/(total

mass of sample). If the sediment was too coarse to core, we

assigned it a value of 4. We then calculated a mean value

for each site. This index ranged from 0 (if all of the sedi-

ment was finer than 0.25 mm) to 4 (if the sediment was too

coarse to core).

We used three methods to estimate exposure to waves

and currents. First, we deployed a modified version (using

a weaker spring to accommodate lower wave forces) of the

dynamometer of Bell and Denny (1994), which is designed

to measure peak wave forces. Second, we set out clod cards

(Petticrew and Kalff 1991) to get a general estimate of the

intensity of water movement at each site. Clod cards were

*600 g masses of plaster-of-paris molded in beveragecups and attached to cement blocks. After they were

retrieved from the field, they were oven-dried and

Hudson shore zone biodiversity

123

reweighed to calculate mass loss, which is roughly pro-

portional to water movement. We set out three

dynamometers and six clod cards at each site for *10 days.Because we had only a few dynamometers, we deployed

dynamometers and clod cards at the six sites in one reach of

river, then later moved on to the other reaches. Third, we

assumed that sediment grain size would reflect peak physical

forces at a site, and used our estimates of sediment grain size

as one measure of exposure. For the sites in each reach

of river (the reaches were treated individually because

dynamometers and clod cards were deployed at different

times in different reaches), we ranked the dynamometer

readings from 1 (most exposed) to 6 (least exposed),

mass loss of clod cards from 1 to 6, and sediment grain

size from 1 to 6. We then summed these three estimates

of exposure for each site, and ranked them for all sites in

the river from 1 (most exposed site) to 18 (least exposed

site).

We measured local substratum roughness (rugosity)

using the chain technique of Frost et al. (2005). In this

method, a 1-m long chain is placed to conform as closely as

possible to the contours of the bottom and the distance

covered by the chain is measured with a taut tape.

Roughness is the ratio of the chain length (1 m) to the tape-

measured distance. We measured rugosity parallel and

perpendicular to the shoreline at nine places using a chain

with a 12-mm link (low water, mean water, and high water

at upriver end, center, and downriver end of each site).

We estimated the complexity of the shoreline in plan

view by stepping off the length of the 100-m stretch of

shoreline at mean low water level using a pair of 1-m wide

calipers. Complexity is estimated as the ratio of shoreline

length measured using calipers to that measured using a

taut line (i.e., 100 m). Values for complexity range from

one for perfectly straight shorelines to higher values for

complex shorelines.

In addition, we tested the effect of nearby beds of sub-

mersed aquatic vegetation (SAV) (chiefly Vallisneria

americana). SAV beds could serve as a source of wrack,

which is important to shore zone invertebrates and plants

(e.g., Backlund 1945; Minchinton 2002), as well as serving

directly as a source of fish and invertebrate colonists to

nearby shore zones. We constructed an SAV index as the

area of the nearest mapped bed of SAV (in m2) divided by

the squared distance to the bed (in m), using data described

by Nieder et al. (2004).

Biological sampling

We sampled fishes, plants, aquatic invertebrates, terrestrial

invertebrates, and snails living in each shore zone site. We

sampled fishes by electrofishing using a Smith-Root Type

VI-A electrofisher mounted in a 6.4-m long john boat. Sites

were continuously shocked until we had covered the length

of the study site or 5 min total of shocking time had

elapsed, whichever occurred first. Sampling occurred

within 3 h of high tide. Fishes were dipped and placed in

an onboard live well. After we finished sampling, we

identified, measured, and counted fish. If more than 25

individuals of a species were collected at a site, the first 25

individuals were measured. We sampled fish three times:

spring, summer, and autumn of 2008.

We sampled terrestrial plants using timed searches on

420 August 2008. Timed searches are an efficient way to

locate rare species (cf. Strayer and Smith 2003). We

walked each site for three consecutive 10-min periods (a

total of 30 min searching) and recorded as many plant

species as possible between the low water mark and 1 m

above the mean high water mark. This search was not

sufficient to detect all plant species, especially at high-

diversity sites. As a result, our data probably underestimate

differences between high- and low-diversity sites. Never-

theless, inspection of the species-accumulation curves

suggested that this method detected [75% of the vascularplant species at each site.

We sampled aquatic invertebrates using two methods.

We took kick samples at all sites, using a 0.5-mm mesh

D-net for 10 min/site, deliberately attempting to sample the

full range of available habitats. We also sampled benthic

invertebrates using 5-cm diameter hand-held cores. We

tried to take samples at the upriver end, middle, and

downriver end of each site at depths of 0.33 and 0.67 m

below the low-water mark. Substrata were too coarse to

allow such core samples to be taken at all sampling loca-

tions. These core samples were sieved in the field through a

0.5-mm mesh sieve. All aquatic invertebrate samples were

preserved in the field in 10% buffered formalin and sorted

in the laboratory under 69 magnification, and identified to

the lowest practical taxonomic level (usually classes or

orders). We took samples of aquatic invertebrates between

5 May6 June and again in 420 August 2008.

We sampled terrestrial invertebrates using a backpack

aspirator. We spent 10 min at each site taking a sample

from the ground and an additional 10 min sampling the air

and vegetation, again attempting to cover the full range of

microhabitats available at each site. Invertebrates were

killed in the field using a killing jar with ethyl acetate, then

stored in the freezer until they were identified to the lowest

practical taxonomic level (usually orders or families). We

took samples of terrestrial invertebrates in early June and

again in early September 2008.

We examined land and aquatic gastropod communities

in detail, identifying all specimens to genus or species. We

sampled aquatic gastropods by taking a 30 s sweep of 1 m2

of substrate at a depth of 1 m below the low tide mark with

a 1.2-mm mesh D-net. Intertidal gastropods were sampled

D. L. Strayer et al.

123

by handpicking and washing substrate from a 1 m2 plot at

mid-tide level through a 1.4-mm mesh sieve. Upland gas-

tropods were collected within 10 m (horizontally) inland of

the high tide mark, by handpicking and sieving soil through

a 1.4-mm mesh sieve. Each type of sample was repeated

three times, near the upriver end, middle, and downriver

end of each site. We collected gastropods twice at all sites,

on 13 June and 29 June2 July 2008. All gastropods

collected were preserved in the field in 95% ethanol, and

identified in the laboratory using Pilsbry (19391948);

Burch (1962, 1989); Harman and Berg (1971); Strayer

(1990); and Jokinen (1992).

Statistical analyses

We conducted four statistical analyses for most taxonomic

groups. First, we used one-way ANOVA (following data

transformation, where needed) to test for differences in

abundance or diversity across the six types of shore zones.

We used Gini diversity as a measure of taxonomic diversity

because it is relatively insensitive to sample size (Gotelli

and Ellison 2004). Second, we ran two-dimensional

nonmetric multidimensional scaling ordinations (NMS,

PC-ORD 5.10) to display variation in community compo-

sition across sampling sites (McCune and Grace 2002). We

transformed the data using a log10 (X ? 1) transformation

and removed taxa present at fewer than three sites prior to

ordination. Plant ordinations were based on presence-

absence data, because we did not collect data on relative

abundance of plants. We then used multi-response per-

mutation procedures (MRPP) to test for significant

differences in community composition among shore types.

Third, we used multiple regression using Pearson correla-

tions (Table 4) and best subset regression in Statistix 9.0 to

identify associations between biodiversity (abundance and

species richness of each taxon, scores on ordination axes)

and the physical characteristics of the study sites. We

transformed shore zone width using a log10 (X ? 0.1)

transformation and the SAV index using a log transfor-

mation prior to statistical analyses.

Results

Physical properties of Hudson River shore zones

The shore zones that we studied were highly varied, both

within and among shore types (Table 1). Important envi-

ronmental characteristics differed significantly among

shore types, as well as between natural and engineered

shores (Table 2). The primary axis of variation in the

physical characteristics of shore zones, which accounted

for 61% of variance in a principal components analysis

(Fig. 2), distinguished flat, sheltered, fine-grained shores

(e.g., sand) from steep, exposed, coarse-grained shores

(e.g., bulkheads). A secondary axis (accounting for an

additional 28% of variance) separated sites with complex

shorelines and rough surfaces from those with simple

shorelines and smooth surfaces.

Terrestrial plants

We identified 190 taxa of plants in Hudson River shore

zones, 46 of which are alien to the region. Plants that were

especially frequent (detected at [50% of sites) includedbush honeysuckles (Lonicera spp.), false indigo (Amorpha

Table 1 Environmental characteristics (mean, with range in parentheses) of Hudson River shore zones

Shore

type

Shore zone

width* (m)

Pebble count* Diversity of

pebble count

Shoreline

complexity

Roughness* Exposure* Sediment

organic

matter (%)

Subtidal

sediment index

Sand 124 (55257) 1.0 (0.961.1) 0.25 (0.030.65) 1.03 (1.001.06) 1.06 (1.021.10) 16 (1418) 2.2 (1.52.8) 0.47 (0.140.79)

Rock 24 (8.952) 3.4 (1.64.8) 0.68 (0.630.76) 1.08 (1.001.12) 1.27 (1.121.40) 12 (717) 3.7 (2.35.3) 2.0 (0.063.3)

Bedrock 30 (8.372) 4.7 (3.66.2) 0.66 (0.330.85) 1.11 (1.031.20) 1.19 (1.171.21) 9.8 (414) 3.2 (3.03.5) 2.4 (1.03.4)

Riprap 8.4 (6.211) 5.0 (4.55.6) 0.59 (0.450.73) 1.23 (1.111.44) 1.45 (1.301.54) 8 (610) 2.9 (0.644.2) 2.5 (1.23.8)

Cribbing 20 (052) 4.2 (2.75.7) 0.71 (0.650.77) 1.06 (1.001.11) 1.37 (1.181.58) 10 (714) 2.7 (2.33.2) 1.6 (0.522.6)

Bulkhead 0.8 (02.4) 6.4 (5.37) 0 (00.45) 1 (1.001.00) 1.15 (1.001.24) 2.7 (15) 2.7 (2.72.7) 1.3 (0.831.7)

Variables marked with asterisks differed significantly (ANOVA, p \ 0.05) among shore types

Table 2 Differences between natural and engineered shore zonesalong the Hudson River (n = 9 for each kind of shore zone)

Variable Natural Engineered p

Log (shore zone width) (m) 1.51 0.68 0.006

Exposure 12.6 6.9 0.01

Rugosity 1.17 1.32 0.07

Pebble count 3.1 5.2 0.02

Native plant richness 30.8 14.8 0.0005

Total plant richness 49.7 34.2 0.01

% alien plants 14.9 27.1 0.002

Abundance of small fish

(square-root transformed)

8.2 3.4 0.04

We show only variables for which significant (p \ 0.1) differencesexist between natural and engineered shore zones

Hudson shore zone biodiversity

123

fruticosa), Virginia creeper (Parthenocissus quinquefolia),

wild grapes (Vitis spp.), dogwoods (Cornus spp.), ashes

(Fraxinus spp.), cottonwood (Populus deltoides), golden-

rods (Solidago spp.), sugar maple (Acer saccharum),

poison ivy (Toxicodendron radicans), bedstraw (Galium

spp.), roses (Rosa spp.), and elms (Ulmus spp.).

Total taxonomic richness of plants did not vary sig-

nificantly among shore types (p = 0.55), but was higher

along natural than engineered shores (Table 2). Species

richness of alien plant species tended to be higher along

engineered shores than natural shores (Fig. 3), but this

difference was not significant (p = 0.43). Species rich-

ness of native plants differed marginally among shore

types (Fig. 3, p = 0.05), and was higher along natural

than engineered shores (Table 2). As a consequence of

opposing trends in native and alien plant species, the

percentage of shore zone plants that are aliens differed

significantly among shore types (Fig. 3, p = 0.0003) and

was significantly higher along engineered shores than

natural shores (Table 2).

The ordination (Fig. 5) shows some clustering by shore

type (e.g., the mowed bulkhead sites grouping in the lower

left), but with considerable overlap among shore types, so

that overall this clustering was not statistically significant

(p = 0.18, MRPP). Axis 2 was significantly (p \ 0.05)correlated with exposure (r2 = 0.45). Sites near the bottom

end of axis 2 were exposed, narrow shore zones with coarse

sediments, while sites near the top end of the axis were

sheltered, wide, and fine-grained. Typical plants of sites

near the lower left of the ordination were lawn weeds such

as butter-and-eggs (Linaria vulgaris), English plantain

(Plantago lanceolata), chicory (Cichorium intybus), and

bedstraw (Galium spp.), while shrubs such as willows

(Salix spp.), witch hazel (Hamamelis virginiana), and

ninebark (Physocarpus opulifolius) were typical of sites

near upper right. Axis 1 was not significantly correlated

with any of the environmental variables that we measured.

Attempts to predict other characteristics of the shore

zone plant community from the physical variables we

measured were only partly successful. Models to predict

native, alien, or total species richness from physical vari-

ables had R2 \ 0.25. Echoing the differences we sawacross shore types, the percentage of shore zone plants that

were aliens was strongly (r2 = 0.50) and positively cor-

related with exposure (Fig. 4).

Fishes

We collected 1121 individual fishes, representing 24 species.

Numerical dominants included spottail shiner (Notropis

hudsonius, 41%), banded killifish (Fundulus diaphanus,

25%), pumpkinseed (Lepomis gibbosus, 7%), blueback

herring (Alosa aestivalis, 5%), white perch (Morone ameri-

cana, 5%), and alewife (Alosa pseudoharengus, 4%). The

abundance, but not the Gini diversity (p = 0.39) of fishes

varied significantly among shore types (Fig. 3). Natural

shore zones had higher numbers of small fishes than did

engineered shore zones (Table 2). Sand shores supported

dense communities dominated by small fishes such as spot-

tail shiners and banded killifish, whereas riprap and rock

shores supported communities of moderately high density.

Bulkheads and cribbing supported low numbers of fishes,

and especially low numbers of small fishes.

The physical properties of shore zones that we measured

were significantly correlated with fish communities. Gini

diversity was positively related (R2 = 0.43) to high

roughness (p = 0.005) and low shoreline complexity

(p = 0.06), with roughness having by far the largest

influence (Fig. 4; Table 4). Density of small fishes (total

length\20 cm) was positively correlated (R2 = 0.63) candhigh shoreline complexity, or nearly as well (R2 = 0.52)

with wide shore zones (Fig. 4; Table 4). The density of

large fishes was somewhat correlated (R2 = 0.37,

p = 0.013) with coarse underwater sediments.

Fig. 2 PCA analysis of HudsonValley shore zones on the basis

of physical characteristics

(shore zone width, pebble count,

shoreline complexity,

roughness, and exposure). Axes

are scaled according to the

% variance explained (these two

axes account for 90% of total

variance). The first letter of thesite code gives the section of

river (L lower, M middle,U upper), and the last twoletters give the shore type(BH bulkhead, BR bedrock,CR cribbing, RO natural rock,RR riprap, SA sand)

D. L. Strayer et al.

123

The ordination (Fig. 5) confirmed the existence of clear

differences in fish communities among shore types. The

first ordination axis was significantly (p \ 0.05) correlatedwith sediment grain size (pebble count), exposure, and

slope, and separated steep, high-energy shores (bulkheads

and cribbing) on the left end of the axis from flat, low-

energy shores (sand) on the right end of the axis. The

second axis was not significantly correlated with any of the

physical variables that we measured, nor did we see any

evidence of consistent differences in physical attributes

among the different geographic sections of the river.

Aquatic macroinvertebrates

The D-net samples were dominated by oligochaetes

(overall mean = 275/sample), chironomids (163/sample),

zebra mussels (102/sample), cladocerans (79/sample),

amphipods (59/sample), gastropods (53/sample), and mites

(44/sample). Neither the overall abundance nor diversity of

D-net invertebrates was related to shore type (p = 0.15 and

0.37, respectively), or differed between natural and engi-

neered shore zones (p [ 0.4). Of the dominant groups justlisted, only the abundance of oligochaetes, zebra mussels,

and cladocerans was related to shore type (Fig. 3). Oligo-

chaetes tended to be less abundant off of high-energy

shorelines (bulkheads and cribbing), zebra mussels were

most abundant on coarse-grained shores (bulkheads, bed-

rock, riprap, and cribbing), and cladocerans were most

abundant on bedrock, sand, and riprap.

Community structure of D-net invertebrates varied

across shore types (Fig. 5). Axis 2 was clearly related to

site exposure, with highly exposed sites having coarse-

grained substrates near the upper end of the axis, and

sheltered sites with fine-grained substrates near the lower

Fig. 3 Summary of differences in biodiversity across types of shorezones in the Hudson River: a species richness of native and alienplants, b percentage of plant species found at each site that are alien,c number of small (\20 cm long) fish caught at each site, d number oflarge ([20 cm long) fish caught at each site, e number of oligochaetescaught per D-net sample, f number of zebra mussels caught per D-netsample, g number of cladocerans caught per D-net sample, h Gini

diversity of upland gastropods, i number of upland gastropodscollected. Error bars show 1 SE, letters in or above bars showdifferences between shore types according to Tukeys Studentized

Range (HSD) test at p = 0.05; the vertical dashed line separatesnatural shore zones on the left from engineered shore zones on the

right. Note the differences in scaling on different y-axes

Hudson shore zone biodiversity

123

end of the axis. Axis 1 scores were not closely correlated

with any of the environmental variables that we measured.

Gini diversity of D-net invertebrates was weakly corre-

lated with shore zone width (i.e., slope) (r = -0.47,

p = 0.05). Generally, abundance of dominant taxa (or of all

taxa combined) and diversity of D-net invertebrates were

poorly correlated with environmental variables (R2 of

models with the lowest AIC usually were\0.2). Exceptionswere models for the abundance of zebra mussels (R2 = 0.66,

positively correlated with pebble count (Fig. 4) and sur-

prisingly negatively correlated with roughness), oligo-

chaetes (R2 = 0.36, positively correlated with SAV index

and exposure), amphipods (R2 = 0.32, positively correlated

with SAV index and pebble count), and gastropods

(R2 = 0.29, positively correlated with pebble count).

We were able to obtain cores at both depths and months

at only 10 of the 18 study sites, so the statistical analyses of

densities and diversity are based on this subset of 10 sites.

Mean invertebrate density was 20,500/m2; dominant

groups were oligochaetes (10,400/m2), chironomids (3,400/

m2), nematodes (1,900/m2), zebra mussels (1,200/m2),

amphipods (660/m2), and sphaeriids (610/m2). Neither total

macroinvertebrate density nor the density of any of the

dominant groups just mentioned varied significantly among

shore types or between natural and engineered shores,

possibly because of the small, unbalanced data set. Fur-

thermore, density of all macroinvertebrates and most

groups taken individually was statistically unrelated to

the environmental characteristics that we measured. The

exception was oligochaetes, whose density was negatively

Exposure index0 2 4 6 8 10 12 14 16 18 20

% o

f pla

nt s

pecie

s th

at a

re a

liens

0

10

20

30

40

50

Substrate roughness0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7

Fish

dive

rsity

0.45

0.50

0.55

0.60

0.65

0.70

0.75

0.80

0.85

Shore zone width +1 (m)1 10 100 1000

Num

ber o

f sm

all f

ish

0

25

100

225

400

Substrate particle size0 1 2 3 4 5 6 7 8

Zebr

a m

usse

ls/D-

net s

ampl

e

0

100

400

900

a

b

c

d

Fig. 4 Examples of relationships between physical characteristicsand biodiversity of Hudson River shore zones: a percentage of plantspecies found at each site that are alien as a function of site exposure

(low values indicate high exposure; r2 = 0.56, p \ 0.001), b fishdiversity (Gini index) as a function of substrate roughness (r2 = 0.25,

p = 0.03), c number of small (\20 cm long) fish caught at each siteas a function of shore zone width (r2 = 0.55, p \ 0.001), d number ofzebra mussels caught per D-net sample as a function of substrate

particle size (larger numbers indicate coarser sediments; r2 = 0.56,p \ 0.001)

D. L. Strayer et al.

123

related to sediment grain size and positively related to the

SAV index (R2 = 0.81, p = 0.003). Gini diversity of sedi-

ment macroinvertebrates in cores did not differ signifi-

cantly across shore types, but was positively correlated

with both shore zone width and site exposure (R2 = 0.67,

p = 0.02; data not shown).

We ordinated samples from each site, depth, and time

separately (we had at least one sample from 14 of the 18

sites). Inspection of this ordination showed that samples

from the two sampling times (May and August) fell out

separately, so we re-analyzed samples from each month

separately. There were no clear relationships between shore

type and the ordination axes for either month (data not

shown), but community structure was clearly related to

environmental characteristics, especially sediment grain

size. In May, axis 1 was significantly related to exposure,

the SAV index, and sediment organic content (R2 = 0.65,

p = 0.0001), while axis 2 was related to sediment grain

size and organic content (R2 = 0.65, p = 0.0001). In

August, axis 2 was unrelated to measured environmental

characteristics, but axis 1 was related to sediment grain size

and organic content, log (shore zone width), and exposure

(R2 = 0.61, p = 0.0009).

Terrestrial invertebrates

Our samples of terrestrial invertebrates were strongly

dominated by dipterans (79% of animals collected, on

Fig. 5 Ordinations of community structure of a fishes, b plants,c aquatic invertebrates (D-net samples), d terrestrial invertebrates.The first letter of the site code gives the section of river (L lower,M middle, U upper), and the last two letters give the shore type (BHbulkhead, BR bedrock, CR cribbing, RO natural rock, RR riprap,

SA sand). The community structure of different shore types differedsignificantly (according to MRPP) for fish (p = 0.001) and aquaticinvertebrates (p = 0.008), but not for plants (p = 0.18) or terrestrialinvertebrates (p = 0.15)

Hudson shore zone biodiversity

123

average) and hemipterans (8%). Other taxa that constituted

[1% of the catch included gastropods, collembolans, spi-ders, mites, ants, and beetles. There were no significant

differences among shore types, or between natural and

engineered shores, in densities of total invertebrates or of

any of the taxa just named, except for mites (p = 0.01), or

in Gini diversity of terrestrial invertebrates.

Density of terrestrial invertebrates was positively related

to the SAV index (R2 = 0.31, p = 0.02), perhaps reflecting

inputs of wrack from nearby SAV beds. Densities of specific

invertebrate groups were related to environmental charac-

teristics as follows (data not shown): mites (complexity[?],

logSAV[?], R2 = 0.44, p = 0.01); beetles (log-shore zone

width[-], pebble count[-], R2 = 0.48, p = 0.008); coll-

embolans (logSAV[?], R2 = 0.31, p = 0.02); dipterans

(logSAV[?], rugosity[-], R2 = 0.45, p = 0.01). Densities

of spiders, ants, gastropods, and hemipterans, and Gini

diversity were not significantly correlated with any of the

environmental features that we measured.

The ordination results separated the sampling sites

partly on the basis of environmental characteristics (the

two bulkhead sites near the top of Fig. 5 both are

mowed lawns) and partly on the basis of geography (note

the cluster of upper-river sites in the lower-right corner of

the figure). Axis 1 scores were weakly related to log

SAV index (R2 = 0.18, p = 0.08), whereas axis 2 scores

were not related to the environmental variables that we

measured.

Gastropods

We found 20 species of freshwater and land gastropods

along the Hudsons shores. The amphibious Fossaria sp.

was most abundant, with the freshwater Physella sp. and

Littoridinops tenuipes also being common and widespread.

Gastropod communities were highly variable across sites,

and neither abundance nor diversity (Gini index) differed

significantly among shore types for subtidal or intertidal

gastropods. There were significant differences across shore

types for upland gastropods, however, with higher densities

and diversities along sandy and rocky shores (Fig. 3).

Ordinations of gastropod communities (not shown) did not

clearly distinguish among shore types, although commu-

nities of flat slopes were weakly separated from those of

steep slopes.

Although there were few clear differences in gastropod

communities among the pre-defined shore types, high

density and diversity of gastropods were consistently

associated with high structural complexity of the shore

zone. Abundance of subtidal gastropods was significantly

correlated (R2 = 0.53, p = 0.005) with high shoreline

complexity, and abundance of intertidal gastropods was

significantly correlated with high roughness (R2 = 0.36,

p = 0.009). Gini diversity of both subtidal and intertidal

gastropods was positively correlated with rugosity

(R2 = 0.24, p = 0.04; R2 = 0.23, p = 0.05, respectively).

In the upland zone, gastropod abundance was not signifi-

cantly correlated with any of the environmental variables

that we measured, but Gini diversity was correlated

(R2 = 0.53, p = 0.004) with pebble count (-) and expo-

sure (?).

Discussion

Hudson River shore zones support rich biological com-

munities. Even in a cursory survey, we found hundreds of

taxa of plants and dozens of species of fishes, as well as

diverse communities of terrestrial and freshwater inverte-

brates. The densities of aquatic invertebrates are at least

twice as high as those in the rest of the river (Strayer and

Smith 2001).

The Hudsons shore zones are enormously varied in

terms of their physical characteristics (Table 1) and the

biodiversity they support (Table 3). Much of this variation

was associated with our classification of shore type, even

though this classification was very simple and based only

on conditions around the high-water line. Presumably, a

more sophisticated classification that considered adjoining

land use (e.g., forest, railroad, lawn) and intertidal and

subtidal conditions (depth, particle size, submersed vege-

tation) more thoroughly would be even more successful in

explaining variation in biodiversity.

As a class, engineered shore zones differed from nat-

ural shore zones (Tables 1, 2). Engineered shores were

narrower (and therefore steeper), coarser-grained, and more

exposed than natural shore zones. Engineered sites may

have been more exposed prior to human activity (i.e.,

defenses were built on the most naturally exposed sites, so

the engineering is a result of the naturally high exposure) or

often because the shoreline was built out into deeper water

or more exposed locations (i.e., the high exposure is a

result of the engineering activity). Surprisingly, some of

the engineered shore zones (riprap and cribbing) were

rougher than many natural shores, at the scale of our

roughness measurements.

Community composition of most taxonomic groups

differed across shore types (Fig. 5). For the majority

(13/18) of our biological attributes, there was some rela-

tionship with a physical variable that we measured

(Table 4). Because we identified invertebrates only to a

coarse taxonomic level, we probably underestimated the

importance of shore type in determining the composition of

the invertebrate community. If we had identified all of the

invertebrates to species or genus, we might have seen

clearer separations among the invertebrate communities of

D. L. Strayer et al.

123

different shore types (e.g., Lenat and Resh 2001; but see

Bailey et al. 2001). Overall, we found that shore zone types

supported quite different biological communities and often

these could be traced to relationships between specific taxa

and some physical attribute of the shore zone. For the

biodiversity variables that differed significantly across

shore types (Table 3), engineered shore zones tended to

have less desirable characteristics than natural shore zones:

fewer fishes, fewer species of native plants, and a higher

percentage of alien plants. It is worth noting that riprapped

shore tended to have more desirable values for biodiver-

sity variables than the other engineered shore types,

Table 3 Summary of differences in major ecological characteristics of different types of shores along the Hudson River

Sand Rock Bedrock Riprap Cribbing Bulkhead p

Fish diversity 77 95 95 100 100 89 0.32

Abundance of large (>20 cm long) fish 6 55 100 61 15 21 0.01

Abundance of small (

presumably a result of its higher physical complexity and

lower wave reflectivity (cf. Pister 2009).

Nevertheless, there was a great deal of variation in

biodiversity among our study sites that was not related to

shore type. Even our relatively small sample (18 sites or

1.8 km out of *300 km of shoreline) revealed large var-iation in both shore zone character (several to 100-fold

differences) and associated biota. Some of this variation

surely arose from unmeasured physical variables and a host

of biotic interactions. Moreover, while it may be tempting

to view relationships between shore zone attributes and

biota as causal the patterns may be driven by indirect

interactions or covariation with other factors. Our findings

have important implications for management or design of

shore zones capable of supporting higher biodiversity but

in our desire to encompass as much of the range in shore

zone function as possible we could not explore these

mechanistic linkages.

Not surprisingly, no single shore type provides high

values of all ecological functions (Table 3). Wide, sandy,

dissipative shores support many small fishes and a high

proportion of native plant species, but few large fishes.

Thus, no shore zone maximizes all kinds of biodiversity.

Our study sites can be broadly arranged according to

their physical characteristics along the two axes of expo-

sure and complexity (Fig. 2). Both of these dominant

physical axes appear to affect shore zone biodiversity.

Community composition varies along the first axis (defined

by the related variables of exposure, sediment grain size,

shore zone slope, and shore zone width), as shown both by

the ordinations and the responses of individual taxa such as

small fishes (which avoid highly exposed shores) and zebra

mussels (which favor highly exposed shores). The associ-

ation between alien plants and high exposure and

engineered shorelines may result from the ability of alien

plants to exploit disturbed sites (e.g., Davis et al. 2000).

The second axis of physical complexity (expressed by

variables such as substrate roughness or shoreline com-

plexity) may generally favor high biodiversity (Fig. 4;

Barwick 2004; Brauns et al. 2007; Paetzold et al. 2008;

Pollock et al. 1998; Sass et al. 2006; Strayer and Findlay

2010). We suggest that these two physical axes (exposure

and complexity) may generally be useful in explaining

variation in predicting the species composition, diversity,

and dominant ecological traits of shore zone communities.

Although we were able to point to the importance of

exposure to shore zone communities, scientists and man-

agers will need to develop faster, easier, more reliable, and

more standardized ways of measuring exposure, which

up to this point has often been treated as a narrative vari-

able or simply assumed to be related to fetch (e.g., Keddy

1982; Brodersen 1995; Ekebom et al. 2003). When we say

exposure, do we mean average wave force, peak wave

force, peak forces of any kind (whether wave or ice-push),

or all of the preceding? How do we reliably measure these

forces so we can compare across studies?

Clearly, our results present only a broad and pre-

liminary assessment of how the biological communities

of shore zones vary according to their physical charac-

teristics and human engineering. Nevertheless, even at

this early stage in our investigations, we can offer the

following suggestions for shore zone management.

Because biotic communities consistently vary along the

gradient between wide, sheltered, fine-grained shore

zones and narrow, exposed, coarse-grained shore zones,

it seems important to preserve or restore a mix of shore

types all along this gradient, at least to the extent that

they occurred naturally. In particular, the common trend

to replace wide, sheltered shore zones with narrow,

exposed shore zones (Strayer and Findlay 2010) may

require special protection for the former to preserve

biodiversity in intensively developed waterways. At a

finer scale, biodiversity often is enhanced by local

roughness and heterogeneity (e.g., Barwick 2004; Brauns

et al. 2007; Paetzold et al. 2008; Pollock et al. 1998;

Strayer and Findlay 2010). Thus, the common tendency

to destroy such local heterogeneity by using uniform

building materials, smooth grades, and straight-line

designs may have strongly negative consequences for

shore zone biodiversity, and should be avoided.

We suggest two lines of future research that should be

useful in achieving the level of understanding needed to

manage shore zone biodiversity well. First, comparative

field investigations like our study might be focused more

tightly on a particular type of shore zones (e.g., riprapped

shores) or a particular taxonomic or functional group. This

tighter focus would allow for use of sampling techniques

that are especially well-suited to that shoreline type or

taxonomic group, sampling designs well-matched to the

taxonomic group (e.g., seasonal or diurnal sampling), and

finer levels of identification or analysis (e.g., body size or

condition). Ultimately, this tighter focus should lead to

higher resolution, lower variance, and increased statistical

power to detect differences among shore types or along

environmental gradients of interest.

Second, it would be very valuable to conduct large-scale

field experiments to test how factors that we have identified

as potentially important might affect shore zone biodiver-

sity. An experimental approach would allow a much

stronger test of the effects of characteristics that are espe-

cially amenable to management intervention (e.g., size of

riprap, slope of revetments, roughness of bulkheads). As

others have suggested (e.g., Doyle et al. 2008), it may be

possible to include such experiments in the construction of

new shore zone defenses or the repair of aging shore zone

infrastructure.

D. L. Strayer et al.

123

Acknowledgments We thank the Hudson River Foundation, theTibor T. Polgar Fellowship program, and the Cooperative Institute of

Coastal and Estuarine Environmental Technology for financial sup-

port. We appreciate technical assistance from Noah Beck, Zara

Dowling, Deana Grimaldi, Robin Schmidt, and Andy Snyder,

encouragement from Betsy Blair, and very helpful comments from

two anonymous reviewers.

References

Airoldi L, Beck MW (2007) Loss, status and trends for coastal marine

habitats of Europe. Oceanogr Marine Biol Annu Rev

45:345405

Backlund HO (1945) Wrack fauna of Sweden and Finland: ecology

and chorology. Opuscula Entomologica Suppl 5:236 ?6 plates

Bailey RC, Norris RH, Reynoldson TB (2001) Taxonomic resolution

of benthic macro invertebrate communities in bioassessments.

J North Am Benthol Soc 20:280286

Barwick DH (2004) Species richness and centrarchid abundance in

littoral habitats of three southern US reservoirs. North Am J Fish

Manag 24:7681

Bell EC, Denny MW (1994) Quantifying wave exposure: a simple

device for recording maximum velocity and results of its use at

several field sites. J Exp Mar Biol Ecol 181:929

Brauns M, Garcia F, Walz N, Pusch MT (2007) Effects of human

shoreline development on littoral macro invertebrates in lowland

lakes. J Appl Ecol 44:11381144

Brodersen KP (1995) The effect of wind exposure and filamentous

algae on the distribution of surf zone macroinvertebrates in Lake

Esrom, Denmark. Hydrobiologia 297:131148

Burch JB (1962) How to know the eastern land snails. W.C Brown,

Dubuque

Burch JB (1989) North American freshwater snails. Malacological

Publications, Hamburg

Caraco NF, Cole JJ, Raymond PA, Strayer DL, Pace ML, Findlay

SEG, Fischer DT (1997) Zebra mussel invasion in a large, turbid

river: phytoplankton response to increased grazing. Ecology

78:588602

Coupland CT, Duarte CM, Walker DI (2007) High metabolic rates in

beach cast communities. Ecosystems 10:13411350

Davis MA, Grime JP, Thompson K (2000) Fluctuating resources in

plant communities: a general theory of invasibility. J Ecol

88:528534

Doyle MW, Stanley EH, Havlick DG, Kaiser MJ, Steinbach G, Graf

WL, Galloway GE, Riggsbee JA (2008) Aging infrastructure and

ecosystem restoration. Science 319:286287

Ekebom J, Laihonen P, Suominen T (2003) A GIS-based step-wise

procedure for assessing physical exposure in fragmented archi-

pelagos. Estuar Coast Shelf Sci 57:887898

Frost NM, Burrows MT, Johnson MP, Hanley ME, Hawkins SJ

(2005) Measuring surface complexity in ecological studies.

Limnol Oceanogr Methods 3:203210

Geyer WR, Chant R (2006) The physical oceanography processes in

the Hudson River estuary. In: Levinton JS, Waldman JR (eds)

The Hudson River estuary. Cambridge University Press, New

York, pp 2438

Gotelli NJ, Ellison AM (2004) A primer of ecological statistics.

Sinauer, Sunderland

Harman WN, Berg CO (1971) The freshwater snails of central

New York with illustrated keys to the genera and species.

Search (Cornell University Agricultural Experiment Station)

1(4):168

Jokinen EH (1992) The freshwater snails (Mollusca: Gastropoda) of

New York state. NY State Mus Bull 482:1112

Keddy PA (1982) Quantifying within-lake gradients of wave energy:

interrelationships of wave energy, substrate particle size and

shoreline plants in Axe Lake, Ontario. Aquat Bot 14:4158

Kennedy TL, Turner TF (2011) River channelization reduces nutrient

flow and macroinvertebrate diversity at the aquatic terrestrial

transition zone. Ecosphere 2 (art35). doi:10.1890/ES11-00047.1

Lazorchak JM, Klemm DJ, Peck DV (eds) (1998) Environmental

Monitoring and Assessment Programsurface waters: field

operations and methods for measuring the ecological condition

of wadeable streams. EPA/620/R-94/004F. United States Envi-

ronmental Protection Agency, Washington, DC

Lenat DR, Resh VH (2001) Taxonomy and stream ecologythe

benefits of genus- and species-level identification. J North Am

Benthol Soc 20:287298

McCune B, Grace JB (2002) Analysis of ecological communities.

MjM Software Design, Gleneden Beach

McLachlan A, Brown AC (2006) The ecology of sandy shores, 2nd

edn. Academic Press, San Diego 392 pp

Minchinton TE (2002) Disturbance by wrack facilitates spread of

Phragmites australis in a coastal marsh. J Exp Mar Biol Ecol281:89107

National Research Council (2007) Mitigating shore erosion along

sheltered coasts. The National Academies Press, Washington

DC, 174 p

Nieder WC, Barnaba E, Findlay SEG, Hoskins S, Holochuck N, Blair

EA (2004) Distribution and abundance of submerged aquatic

vegetation and Trapa natans in the Hudson River estuary.J Coast Res 45:150161

Paetzold A, Yoshimura C, Tockner K (2008) Riparian arthropod

responses to flow regulation and river channelization. J Appl

Ecol 45:894903

Petticrew EL, Kalff J (1991) Calibration of a gypsum source for fresh-

water flow measurements. Can J Fish Aquat Sci 48:12441249

Pilsbry HA (19391948) The land Mollusca of North America (north

of Mexico). Academy of natural sciences at Philadelphia

monograph 3, vols 1 and 2

Pister B (2009) Urban marine ecology in southern California: the

ability of riprap structures to serve as rocky intertidal habitat.

Mar Biol 156:861873

Polis GA, Hurd SD (1996) Linking marine and terrestrial food webs:

allochthonous input from the ocean supports high secondary

productivity on small islands and coastal land communities. Am

Nat 147:396423

Pollock MM, Naiman RJ, Hanley TA (1998) Plant species richness in

riparian wetlands: a test of biodiversity theory. Ecology

79:94105

Sabo JL, Sponseller R, Dixon M, Gade K, Harms T, Heffernan J, Jani

A, Katz G, Soykan C, Watts J, Welter J (2005) Riparian zones

increase regional species richness by harboring different, not

more, species. Ecology 86:5662

Sass GG, Kitchel JF, Carpenter SR, Hrabik TR, Marburg AE, Turner

MG (2006) Fish community and food web responses to a whole-

lake removal of coarse woody habitat. Fisheries 31:321330

Scholten M, Anlauf A, B}uchele B, Faulhaber P, Henle K, Kofalk S,Leyer I, Meyerhoff J, Purps J, Rast G, Scholz M (2005) The

River Elbe in Germanypresent state, conflicting goals, and

perspectives of rehabilitation. Archiv f}ur Hydrobiologie Supple-mentband 155:579602

Simpson KW, Fagnani JP, DeNicola DM, Bode RW (1985)

Widespread distribution of some estuarine crustaceans (Cyath-ura polita, Chiridotea almyra, Almyracuma proximoculi) in thelimnetic zone of the lower Hudson River, New York. Estuaries

8:373380

Simpson HJ, Chillrud SN, Bopp RF, Schuster E, Chaky DA (2006)

Major ion geochemistry and drinking water supply issues in the

Hudson River basin. In: Levinton JS, Waldman JR (eds) The

Hudson shore zone biodiversity

123

Hudson River estuary. Cambridge University Press, New York,

pp 7996

Strayer DL (1990) Freshwater Mollusca. In: Peckarsky BL, Fraissinet

PR, Penton MA, Conklin DJ (eds) Freshwater macroinverte-

brates of northeastern North America. Cornell University Press,

Ithaca, pp 335372

Strayer DL, Findlay SEG (2010) Ecology of freshwater shore zones.

Aquat Sci 72:127163

Strayer DL, Smith LC (2001) The zoobenthos of the freshwater tidal

Hudson River and its response to the zebra mussel (Dreissenapolymorpha) invasion. Arch Hydrobiol Suppl 139:152

Strayer DL, Smith DR (2003) A guide to sampling freshwater mussel

populations. Am Fish Soc Monogr 8:1103

Tockner K, Stanford JA (2002) Riverine floodplains: present state and

future trends. Environ Conserv 29:308330

Ward JV, Tockner K, Schiemer F (1999) Biodiversity of floodplain

river ecosystems: ecotones and connectivity. Regul Rivers Res

Manage 15:125139

Wetzel RG (1990) Land-water interfaces: metabolic and limnological

regulators. Verhandlungen der Internationale Vereinigung f}urTheoretische und Angewandte Limnologie 24:624

D. L. Strayer et al.

123

Biodiversity in Hudson River shore zones: influence of shoreline type and physical structureAbstractIntroductionMethodsThe study areaStudy designPhysical variablesBiological samplingStatistical analyses

ResultsPhysical properties of Hudson River shore zonesTerrestrial plantsFishesAquatic macroinvertebratesTerrestrial invertebratesGastropods

DiscussionAcknowledgmentsReferences